AI Agents, Biocomputers, and the Data Center Backlash Reshaping Tech
Updated on November 05, 20265 min read
AI agents were supposed to change everything. This week, Uber revealed that almost nobody is actually using them, while Meta quietly snapped up a startup betting that multi-agent coding is the future. These two data points sit side by side in a week full of contradictions worth paying attention to.
AI agents: the gap between hype and reality
Uber disclosed that during its busiest week last quarter, just four people booked a ride through an AI agent. Four. That number is a useful anchor when you hear bold claims about autonomous AI assistants taking over everyday tasks. Read Uber's AI agent usage figures on TechCrunch.
Meanwhile, Meta moved in the opposite direction, acquiring AI coding startup Conductor, whose product lets engineers spin up teams of agents that tackle different parts of a software project at the same time. The gap between consumer AI agents (barely used) and developer-focused AI agents (heavily invested in) is worth keeping in mind if you are thinking about where the near-term opportunities actually are. Details on Meta's Conductor acquisition via TechCrunch.
Apple tightens its AI spending
Apple is reportedly cutting internal AI budgets for its engineering teams. The company has not explained why publicly, but the move is notable given how much pressure Apple has faced to demonstrate meaningful AI progress across its products. For aspiring developers, it is a reminder that "AI investment" is not uniform across the industry. Some companies are pulling back internally even while ramping up externally-facing features. More on Apple's AI budget cuts from TechCrunch.
Separately, Apple announced that starting with iOS 27.3, it will scan iCloud links shared publicly for harmful content. The change applies to any item set to "Anyone with the Link" access. Privacy advocates will have questions, and if you build apps that use iCloud sharing, this is a policy shift to track. Apple's iCloud content scanning explained by The Verge.
Data centers face a public backlash
Voters across the US pushed back hard against data center expansion in midterm ballot initiatives, with backers of new facilities losing nearly every contest they entered. The opposition cut across party lines, driven largely by concerns about energy use, water consumption, and local infrastructure strain. Big Tech responded by framing the backlash as political rather than a genuine rejection of the industry's expansion plans, an interpretation that many observers found unconvincing. CNET's breakdown of the data center ballot results and TechCrunch's analysis of what comes next are both worth reading together.
For anyone studying cloud computing or infrastructure, this is a live debate about where the physical layer of the internet gets built and who gets a say. It is not just a policy story.
Qualcomm and AMD bet on data centers as chips shift
While voters resist new data centers, the chip companies supplying them are thriving. AMD posted a record quarter at $13.4 billion in revenue, with its data center business doubling year over year. Gaming revenue, by contrast, fell sharply. The story shows how quickly revenue mix can shift inside a large hardware company when one market accelerates and another contracts. Ars Technica's AMD earnings breakdown.
Qualcomm told a similar story. With smartphone sales dropping significantly compared to last year, the company laid out a plan to grow its data center business through partnerships with Amazon and Meta. If you have been learning about mobile hardware, this pivot is a signal that the chip design skills valued in mobile are increasingly being pointed at server and AI workloads. Qualcomm's data center strategy on The Verge.
Scientists build a computer powered by bacteria
Researchers have built a biocomputer that uses living bacteria, specifically E. coli and V. natriegens, to perform calculations on hard computational problems. This is not a replacement for your laptop. It is early-stage science exploring whether biological systems can handle certain classes of problems that silicon processors struggle with efficiently. The fact that it works at all is the news. CNET's explainer on the bacteria-powered biocomputer.
Two security stories you should not ignore
A group of hackers, believed to be linked to the Kremlin, exploited vulnerabilities in both iOS and Android to target phones belonging to diplomatic figures. The attack used flaws in the operating systems themselves, not just phishing links, which makes it harder for ordinary users to defend against. For anyone studying cybersecurity, this is a case study in how state-level actors operate differently from typical criminal hackers. Ars Technica's report on the iOS and Android exploits.
Separately, the web hosting provider DreamHost disclosed that attackers had exploited a vulnerability to access customer emails and files over several months before the breach was caught. Long-running intrusions like this are common and underscore why monitoring and anomaly detection matter as much as preventing initial access. TechCrunch's coverage of the DreamHost breach.
Nothing teases a screen-free AI device
Startup Nothing, known for its design-forward phones and earbuds, confirmed that its first standalone AI device is coming in December. CEO Carl Pei described it as screen-free and built around productivity, though specifics are thin. The device appears aimed at users who want AI assistance without staring at another display. Whether the product delivers on that premise is an open question, but it represents a genuine attempt to rethink the form factor rather than just adding a chatbot to a phone. The Verge's coverage of Nothing's upcoming device and TechCrunch's report from Nothing's CEO have complementary details.
Open source AI is growing, and so are the business questions
Hugging Face weighed in on a question that matters a lot for the AI industry: as open source models become more capable and widely used by enterprises and governments, how do the companies releasing them actually make money? The business model tension between openness and commercial value is not resolved, and with Chinese open models gaining ground internationally, the competitive stakes are rising. TechCrunch's discussion with Hugging Face on open source AI economics.
The week's stories collectively point toward a tech industry under pressure from multiple directions at once: voters, regulators, tight hardware supply chains, and the basic question of whether AI tools are genuinely useful yet. The answers will shape which skills, products, and business models hold up over the next few years, which makes right now a genuinely interesting time to be learning.

Learn practical tech skills online with Code Labs Academy
Join a supportive community, build real-world skills, and upskill for your next opportunity.
Apply now